Picture two practices on the same street. One ships the foundation — schema, GBP, reviews, FAQ — in a weekend, mostly free. The other spends five figures on an agency that delivers a 40-page AEO strategy deck and zero verified recommendations. The difference is not budget. It is that decks do not change AI answers; complete, consistent, current business data does — and the only way to know it worked is to measure the answers from real devices.
This guide is the foundational layer: what to build yourself, with free and under-$500/mo tools, in roughly a weekend. It is not a pitch for Top3. It is the prerequisite work you need done before Top3 (or anything else) can move the needle. Skip these and any ranking service — ours or a competitor's — is building on sand.
Read this end-to-end, then run a free AI Visibility Check to see where you currently stand before you begin. The audit establishes a baseline — so when you ship your stack, you can measure whether it moved the needle.
What "Minimum Viable" Actually Means
Most AEO content online is either (a) a six-figure enterprise roadmap that pretends to be SMB-friendly, or (b) a generic SEO checklist with the word "AI" sprinkled in. Neither is what a real SMB needs.
Minimum viable AEO has one job: make it possible for an AI engine to name your business as a correct answer to a buyer question. That requires five things — no more, no less — and none of them are glamorous. This is the foundation layer. It makes you eligible to be recommended. Whether you actually get recommended, and at what frequency, depends on how your whole body of work stacks up against your competitors' (covered in How to Rank in ChatGPT) — and you only know by measuring.
The mental model: think of your business like a job candidate. Schema and GBP are your resume — they make you legible to the hiring algorithm. Reviews and entity alignment are your references — they confirm your resume is true. Consistency over time is the actual job interview — the part where most candidates quietly fail. DIY gets you to the interview. Top3 keeps showing up to it every week, and films the verdict.
"We did all the schema stuff, got to 150 Google reviews, rebuilt our GBP. Took 90 days. ChatGPT still named our competitor. Then we called Top3 and the recommendations flipped inside 11 days." — Dental practice owner, Austin
The 5 Pieces Every SMB Needs
These five layers are non-negotiable. Skip any of them and you are fighting with one hand tied. Ship all five and you have a defensible foundation — one that Top3 or any ranking service can build on.
1. Schema Markup (LocalBusiness, Service, Organization, FAQPage, Review)
Schema.org markup is the machine-readable description of your business that AI engines parse before they parse your prose. Without it, the AI is guessing at what you do, where you do it, and who you serve. With it, the AI has a structured, unambiguous answer to every question it might get asked.
The five schema types every SMB should deploy, in order of impact: LocalBusiness (or a more specific subtype like Dentist, Plumber, LawFirm), Service for each service you offer, Organization for your parent entity, FAQPage for your category FAQ, and Review aggregated from your collected reviews. We have a complete copy-paste-ready guide at Schema.org for Small Businesses — every JSON-LD snippet, documented, ready to drop in.
Time to implement: 2–4 hours if you or a developer can edit your site's `<head>`. Free with any CMS. Validation: run every snippet through Google's Rich Results Test before deploying.
2. Google Business Profile — Completed, Not Just Claimed
Every SMB claims GBP. Most do not complete it. The difference between "claimed" and "completed" is about 80% of the recommendation value AI engines derive from GBP. Completed means: full primary and secondary categories, service list with descriptions, service area polygons, hours including holiday hours, attributes (wheelchair-accessible, free wifi, whatever applies), 10+ photos across interior/exterior/team/product, and a business description written in 700 characters that uses the questions buyers actually ask in AI prompts.
The description is where most SMBs fail. Do not write it for Google. Write it for ChatGPT. Use the language a buyer would use when asking an AI engine: "the best X in Y" language, not SEO-keyword-stuffed language. If a prospect would ask "is there a family dentist in Austin who takes Delta Dental?" your GBP description should name that exact phrasing back.
Time to implement: 3–5 hours for a full audit and rewrite. Free. Validation: check the profile completeness score inside GBP and target 100%.
3. Active Review Pipeline — 50+ Reviews, 4.6+ Average, Monthly Velocity
Reviews are not just a trust signal — they are the primary source AI engines use to synthesize what people actually think about you. An AI engine asked "is X any good?" will often quote or paraphrase review content directly. If your review corpus is thin or dated, the AI has nothing to recommend except competitors' richer ones.
The target: 50+ reviews minimum on Google, 4.6+ average rating, and fresh velocity (at least 2–3 new reviews per month). A pile of 200 reviews from 2021 helps less than 40 reviews with five new ones in the last 90 days — AI engines weight recency heavily.
How: set up an automated review-request flow that fires after every completed service/purchase. Tools like NiceJob, Podium, Birdeye, or a simple Zapier chain from your CRM to an email template will all work. Budget: $50–$200/mo depending on tool. Time: 2–3 hours to set up the flow, then automatic.
4. Category FAQ Page — Answer Shaped
AI engines pull heavily from FAQ-style content because it matches the question-shape of how buyers actually prompt them. A well-built category FAQ page gives the AI a ready-to-quote answer for every common buyer question in your vertical.
Structure: 10–20 question/answer pairs, each question phrased exactly as a buyer would type it into ChatGPT, each answer 50–150 words in conversational prose. Include pricing ranges, time-to-completion estimates, and specific differentiators. Wrap the whole page in FAQPage schema (see point 1).
The questions: do not invent them. Use three sources — (1) your actual customer emails and call logs, (2) the "People also ask" box on Google for your category + city, (3) ChatGPT itself — ask it "what are the top 20 questions people ask about X in Y city?" and edit from there. Time: 4–6 hours for a complete first version. Free.
5. Entity Alignment Across Directories and Listings Sites
This is the plumbing nobody talks about. Your business has a presence across dozens of directories and recommendation sites — Yelp, Apple Maps, Bing Places, industry-specific directories (Avvo for lawyers, Healthgrades for medical, Angi for services), and legacy data aggregators (Neustar, Factual, Foursquare). When these entries conflict — different addresses, different phone numbers, different hours — AI engines lose confidence in your entity and demote you.
The fix: run a listings audit (BrightLocal and Whitespark both offer one under $50). Identify inconsistencies. Fix them — either manually if it is 10–15 listings, or via a listings management service (Yext, BrightLocal, Whitespark Local Listings Builder) if it is 40+ listings. Budget: $100–$300/mo if you use a managed service.
Time to implement: 6–10 hours for a full manual fix. Or: this is one of the layers our ops team handles for you inside every Top3 engagement — skip this entirely if you are already a customer.
The Under-$500/mo Stack, Priced
Here is the concrete toolchain. These are not affiliate links — they are the tools we see working consistently in SMB engagements. You can swap any one for a close competitor (most categories have several viable options at similar price points).
Schema markup: Free. Hand-code JSON-LD from our schema guide, or use the free plugin for your CMS (Yoast, Rank Math, Squarespace built-in, Shopify JSON-LD app).
Google Business Profile: Free. Access via business.google.com.
Review automation: $49–$199/mo. NiceJob ($75/mo), Podium ($249/mo, more than needed at MVP), Birdeye ($299/mo — skip at MVP), or Zapier + your email tool ($30/mo for bare minimum).
Category FAQ page: Free. Built into your CMS. ChatGPT or Claude ($20/mo) helps draft the questions list.
Listings audit + fix: BrightLocal ($39/mo for Single Location), Whitespark ($29/mo). Pick one.
AI visibility baseline measurement: Free AI Visibility Check from Top3 — establishes your day-zero baseline so you can measure lift as you ship the stack.
Total monthly cost: $117–$247/mo depending on choices. Total time to ship: a focused weekend if you are technical, or about three weekends if you are not.
What the DIY Stack Can't Do
Here is the honest ceiling of the DIY stack: it makes you eligible. It does not make you chosen. And the gap between eligible and chosen is where every SMB loses to better-funded competitors — unless something closes that gap.
AI engines choose one business from a field of legible candidates based on everything they can read — and how well it agrees with itself. Entity data, local listings, review volume and language, answer-shaped content, recency. When five businesses are all plausible, the one whose footprint is deepest, most consistent, and most current is usually the one that gets named.
The DIY stack builds that footprint once. What it cannot do is hold it: data drifts, engines update, competitors keep working, and a weekend of setup decays into last year's snapshot. And the DIY stack cannot verify itself — engine answers vary by location and session, so checking from your own laptop proves nothing about what your buyers see.
This is why Top3 exists and why we built a 10,000+ real-device farm — the work, done continuously on your behalf, and the measurement that proves it landed, photographed daily from real devices in your market. A DIY stack plus Top3 is the full system. A DIY stack alone leaves maintenance and verification — the deciding factors — to chance.
The DIY stack and Top3 are not alternatives. They are layers of the same system. Ship the DIY foundation first. Then call us when you want it maintained, deepened, and proven — without doing it yourself.
When to Stop DIY and Hire Top3
There is a clean decision rule. Ship the five foundational pieces, let the stack run for 60–90 days, then measure. You are ready to hire Top3 when at least two of these three signals are true:
- AI visibility is flat. A repeat free visibility check shows no movement — buyer queries in your category still do not recommend you by name, 90 days after shipping the stack.
- Competitors are getting recommended. When you ask ChatGPT "best X in Y," competitors you consider peers get named — and a review of their GBP and site does not show material differences from yours. The real differences are usually below the surface — and unverifiable from your laptop.
- AI-referral traffic is zero. Your GA4 shows zero or near-zero referrals from chatgpt.com, perplexity.ai, or gemini.google.com, despite the DIY stack being in place.
At that point, the marginal spend on more DIY work gets you very little. The next dollar of AEO spend is most productive on done-for-you depth plus real-device proof — which is what Top3 is for. Our measured library documents 401 climbs into the top 3 of AI answers across 92 businesses (median 14 days, range 14–48), each with a screenshot receipt. See plans →
Ship the Foundation. Measure. Then Decide.
Everything in this guide is achievable in a weekend for a technical founder, or about three weekends for a non-technical one. Under $250/mo in tools, if you are selective. No agency, no retainer, no 12-month contract. That is what "minimum viable" looks like — and it is where every SMB should start before spending anything on a ranking service.
Start with a free AI Visibility Check to establish your day-zero baseline. Ship the five layers above over the next month. Re-run the check after 60–90 days. If the check shows movement, you have the foundation and should keep operating. If the check is flat and competitors are being named, you are at the DIY ceiling, and handing the work to Top3 is your next move.